Regional scale air quality (AQ) models are typically limited to relatively coarse resolutions when simulating mean pollutant concentrations for each grid cell volume. However, human exposure and risk assessment require more detailed information on the location and magnitude of hazardous air pollutant, or air toxics, concentrations, with particular interest in capturing extreme values, or 'hot spots.' Computational fluid dynamics and coupled large-eddy simulation with photochemistry techniques allow AQ simulations with much finer grid spacings, but these types of simulations are impractical for long time integrations or operational use. Thus, some procedure is needed to represent the subgrid pollutant concentration extremes in regional models without requiring concurrent fine resolution simulations.